Participation à DCMIP 2012 GEM Lat-Lon GEM Yin-Yang€¦ · 3D Advection: Test 11 (3D deformational flow) Correlated tracers in a reversing sheared flow. Q1. LL YY Q1 Test 11 Near

Participation à DCMIP 2012

GEM Lat-Lon versus GEM Yin-Yang

Abdessamad QaddouriVivian LeeMonique TanguayClaude Girard

Collaboration: Bernard Dugas, Ayrton Zadra et Paul Vaillancourt

DCMIP = Dynamical Core Model Intercomparison Project

Droits d’auteurs: Plusieurs éléments de la présentation sont extraits du site DCMIP 2012

3D Advection: Test 11 (3D deformational flow)

Correlated tracers in a reversing sheared flow

Q1

LL YY

Test 11 Near 4900 m t= 6 days Q1

Tracer q is stretched thin after 12 h, but it is still resolved,comes back to its original position after 1 day

3D Advection: Test 12 (Hadley-cell like circulation)

Animation CAM-FV

LL YY

Test 12 medium L60 t= 12 hours

3D Advection: Test 13 Solid body rotation of thin cloud tracer in presence of orography

days HEIGHT levels

Animation GFDL FV3

days MODEL levels

LL YY

Test 13 L60 (Model levels) t=T/2

Evaluate accuracy of the pressure-gradient calculation over a Schaer-type mountain

Test 20 (Steady-State Atmosphere at Rest in Presence of Orography)

LL YY

U

W

LL YY

Test 20 L15 t= 6 days

Scale*10

N-1

N

N-1/2

N+1/2

F

F

N-1

N

σN+1/2

σN-1/2

.

.

( )( )

−∆−=

+=∆=

=∆=

∆+

∆−=

∆

∆+

∆

∆−=

=∆

−−=

∆∆−=

∂∂

∆−+

∆−−=

∆∆−=

∂∂

∂∂−=

∂∂

−−

−+

−

−+

−

−

−−

+

−−

−

−

−+

+−

−−

−

−−

+

++

2/32/

2/2/

1

2/2/1

0 ;21

21

1

1

1

1

1

1

11

21

21

21

21

21

21

21

21

21

21

212

1

212

1

212

1

NNN

NNNN

NNN

N

NN

NN

N

N

N

N

N

NN

NN

NNN

NN

k

kkk

k

kkk

kk

FF

FFF

FF

FFF

Ftt

FF

FFFtF

FFFFFtF

FtF

σσσσδσδσδσδσσσσσδσδσδσδσ

σσσσδσδσδσδσ

σσσσδσδσδσδσ

σσσσδσδσδσδσ

σσσσ

σσσσ

σσσσ

σσσσ

σσσσσσσσ

σσσσσσσσ

σσσσ

σσσσσσσσ

σσσσσσσσ

σσσσσσσσ

σσσσσσσσ

σσσσ

σσσσ

Vertical advection: Eulerian Schemelinear extrapolation near boundaries

is not only possible but in fact effective

Semi-Lagrangian scheme:a similar approach (linear extrapolation)

is not only possible but no doubt desirableIt is now being developed for GEM

=0

σσσσσσσσ

∂∂−=

∂∂ F

tF

LL

U

W

LL

Test 20 L15 t= 6 days

• Corrections à No Top thermodynamic level(Claude et André)

• Extrapolation linéaire• Rcoef (0,2)

Same Scale

Radius of the Earth is scaled so that the simulation is in the non-hydrostatic domain

1

R top) model near the(Spongefriction Rayleigh

Velocity Angular

Step Time

Radius

>

=

Ω=Ω

∆=∆

=

X

RX

XX

tt

X

a a

ref

ref

ref

ref

SMALL PLANET

Test 21: Without Wind Shear (Reducedsize Earth with a circumference at the equator of 80 km)

X=500

Animation MCORE

Test 21 (Flow over Topography, Gravity waves Without Wind Shear)

Reducedsize Earth with a circumference at the equator of 80 km and No rotationx 333.6 m z = 500 m

Schär’s case of Claude (Slice GEM)

Distinguishes between Hydrostatic and NonHydrostatic responses

ENDGAME

Test 21 T’ t=3600 s

LL YYTest 21 T’ t=3600 s

IFS

X=500

Animation MCORE

Test 22 (Flow over Topography, Gravity waves With Wind Shear)

Reducedsize Earth with a circumference at the equator of 80 km and No rotationx 333.6 m z = 500 m

Test 22 T’ t=3600 sLL YY

IFS

X=125

Non-orographic Gravity wavesReducedsize Earth with a circumference at the equator of 320 km and No rotationx 1 km z = 1 km

Animation NIM

Animation MCORE

Example of NonHydrostatic responsein potential temperature ’

Test 30 (Warm Bubble Triggered Gravity Waves)

Test 31 L10 Theta’ t=3600 s

Test 31 L10 Theta’ t=1200 sLL YY

LLTest 31 Theta’ t=1200 s Test 31 Theta’ t=3600 s

ENDGAME

• Corrections à No Top thermodynamic level(Claude et André)

• Extrapolation linéaire

X=1,10,100,1000

Test 41X (Dry Baroclinic Instability)Steady-State Zonal Wind

perturbation

+(20E,40N)

Full time series 1 deg L30 NonHydrostatic over 15 days

Animation IFS

ICON-MPI-DWD

YYLLX=1

X=10

X=100

X=1000

Test 41X Surface pressure t=Day 9

+ Aqua-Planet (Terre couverte d’eau à 29ºC SST)

Test 42 (Moist Baroclinic Instability)

Large-scale condensation in a moist baroclinic waveleads to an intensificationin CAM-FV (1 deg L30), here at Day 9

LL

YY

LL YY

Test 42 1 deg L30 t=Day 9

Test 51 (Tropical Cyclone with Simplified Physics)

Idealized Tropical Cyclone on Aqua-Planet: Simulations with Simplified Physics

100 m

Animation NIM

Evolution of the tropical cyclone for 10 days (L30) Wind speed at the 2nd model level (~150 m)

CAM5-FV .25 deg L30

1. deg L30

Day 3 Day 5 Day 10

LL

YY

.5 deg L30

Day 3 Day 5 Day 10

LL

YY

LL

.25 deg L30 The intensity increases with the resolutionThe cyclone is more concentrated in space with the resolution

Day 3 Day 5 Day 10

YY

Test 52 (Tropical Cyclone with Full Physics) OPTIONAL

Idealized Tropical Cyclone on Aqua-Planet: Simulations with Full Physics .5 deg L30

Animation CAM-SE

SimplePhysics

FullPhysics

CAM5-FV 0.25 L30

Expansion of the cyclone with time when Full Physics

Eye of the cyclone clearly defined

Day 3 Day 5

.5 deg L30

LL

YY

Day 10

Several small cyclones at Day 10 …

SOMMAIRE

• GEM Yin-Yang et GEM Lat-Lon ont participé au projet DCMIP 2012.

• Les testcases comprennent des advections pures, des ondes de gravité orographique et non-orographique sur une planète sans rotation, des simulations Petite Planète pour obtenir un signalNonHydrostatique à faible coût de calcul, des ondes baroclines sèches et humides et des cyclones tropicaux avec physique simplifiée.

• GEM Yin-Yang et GEM Lat-Lon donnent des réponses très semblables.

• Les Test20 et Test31 ont été utiles pour comprendre et corriger une faiblessedans l’advection semi-Lagrangienne près du toit et du sol.

• Pour le Test 52 (Pleine Physique, OPTIONNEL), GEM ne peut maintenir un cyclone unique jusqu’au Jour 10. Un ajustement incomplet d’Aqua-Planèteen lien avec la radiation est probablement en cause.